A stochastic analysis of claim reserving in General Insurance using bootstrapping technique

Abstract

Insurance is one of the key sector in ensuring stability and growth in any economy. Reserving actuaries therefore should do the best they can to ensure that the reserves declared are as accurate as possible. In most insurance companies reserves are calculated using the traditional methods which are based on certain algorithms. These include Chain ladder method, Average cost method,Bornhuetter-Ferguson and Standards method among others. These methods do not take into consideration the actual claim development. Generalised linear models (GLMs) and Bootstrap technique can be used as an alternative as they use the various covariates of the claim process in determining the reserves. In this study I will focus in showing that Bootstrap technique and Generalized Linear Models gives a realistic and structured method of loss reserving. This is because they incorporate more information about the claim process such as Type of claim, Loss, development pattern, Pattern of loss emergence etc. This makes it possible to determine the predictive distribution of the reserve model and to calculate various measures of risk such as the Value at Risk(VaR) at various levels.